Computational Literary Analysis of Tango Lyrics: Meter, Emotions, and Thematic Evolution

Abstract

This interdisciplinary study aims to investigate 6,811 tango lyrics written between 1912 and 1985, organized by lyricist, period, and generation, using computational literary analysis to map their metrical, emotional, and thematic evolution. SICAM was used for automated metrics, NES for sentiment analysis, and LIWC and LDA TF-IDF for topic modeling. The metrical findings indicate that in the early stages of tango, short verses, especially the octosyllable, dominated, reflecting its festive character. Subsequently, complex structures and free verse increased due to the influence of urban modernization. A sustained trend toward greater emotional negativity was confirmed, supporting Borges's thesis that tango “is languishing and growing sadder.” Thematically, throughout its history, the topics we call emotions, death (associated with the loss of “beloved Buenos Aires”), and work increased. The decline over time in the personal theme underscores a progressive increase in abstraction and a less narrative focus in the lyrics. Furthermore, the use of lunfardo (Argentine slang) declined after 1930. This study offers a quantified view of how the aesthetic transformations of tango contribute to constructing the complexities of modernity in the Río de la Plata region.